Identification-robust beta pricing, spanning, mimicking portfolios, and the benchmark neutrality of catastrophe bonds

This paper proposes inference strategies for conditional and unconditional asset pricing models, which cover beta pricing, the zero-beta rate, risk prices, Jensen’s alphas, mean–variance spanning and intersection. We derive analytical conditions that link the identification of the zero-beta rate to spanning. Following finite-sample statistical considerations, the proposed procedures correct for measurement error with mimicking portfolios, and are invariant to fund repacking as well as and robust to: (i) missing factors, (ii) identification of risk price, and (iii) the quality of conditioning information. Empirically, we study benchmark neutrality of Catastrophe bond mutual funds. Results show that potential for diversification is significant before the 2008 financial crisis, yet lack support thereafter.